perl-based language (perl 5.28.1) was used to combine DR-DEGs expression with the clinical data to remove samples with incomplete clinical data and a survival time of 0 or negative. To assess the prognostic value of drug differential genes, a univariate COX risk regression analysis of endometrial carcinoma tumor group data was performed using R package ‘survival’, aiming to screen out drug differential genes significantly associated with survival. To avoid the overfitting problem, a least absolute shrinkage and selection operator (LASSO)-penalized cox regression analysis was performed using R package “glmnet”. Later, a prognostic model was established by multivariate COX risk regression analysis, and two prognosis-related DR-DEGs were finally acquired. Thereafter, the risk score was calculated for each patient using the formula: Risk score = ∑X λ*coef λ. Wherein, X λ stands for the relative expression level of the normalized differential genes for each drug; coef λ for coefficient. Patients in the tumor group were classified into high-risk and low-risk groups based on the median risk score. To determine the role of risk scores in the prognostic model of endometrial carcinoma patients, we performed a separate analysis of overall survival (OS) between high and low risk groups and genes with prognosis-related drug differences, and displayed it with Kaplan–Meier curves. In addition, the “Rtsne” package function of R software was used for principal component analysis (PCA) and t-SNE test. The groups were visualized to explore the distribution of different groups. R package “survivalROC” was used to perform time-dependent receiver operating characteristic (ROC) analysis, so as to test the specificity and sensitivity of the survival prognostic model.
Free full text: Click here
